Military Embedded Systems

Do dragonflies hold the secret to better missile defense?


September 17, 2019

Sally Cole

Senior Editor

Military Embedded Systems

A researcher at Sandia National Laboratories is exploring how dragonfly brains might be wired to be extremely efficient at calculating complex trajectories.

Dragonflies were among the first winged insects to emerge 300 million years ago, predating dinosaurs, and are well- known as excellent fliers and hunters. They’re so adept, as one recent Harvard University study found, that dragonflies caught between 90% to 95% of prey when it was released into an enclosure with the insects.

If you’ve ever looked closely at a dragonfly, its eyes make up most of its head, which gives them amazing vision from nearly every angle. Even when its prey is weaving evasively, a dragon­fly will track it with seemingly instant reflexes; just when the prey thinks it has gotten away, the dragonfly closes in for the kill swiftly from below.

How do they do it? By calculating the distance of their prey, where it’s heading, and the speed at which it’s flying. The dragonfly needs only milliseconds to calculate its optimum angle of approach.

Dragonflies’ combo of quick brains and skilled flying are highly desirable for missile defense systems. To this end, Frances Chance, a computational neuroscientist at Sandia National Laboratories, is trying to figure out whether dragonflies’ tiny brains might be wired for calculating complex trajectories akin to those needed for missile defense. (Figure 1.)

In Sandia computer simulations, faux dragonflies within a simplified virtual environment successfully caught their prey using computer algorithms – designed by Chance – to mimic the way a dragonfly processes visual information while hunting. These positive test results indicate that this programming is fundamentally a sound model, according to Chance.

Her ultimate goal is to determine whether dragonfly-inspired computing can improve missile defense systems, which have the similar task of intercepting an object in flight, by making its onboard computers smaller, but without sacrificing speed or accuracy.

Chance specializes in replicating biological neural networks, basically brains, which require less energy and are better than computers at learning and adapting. Her studies center on neurons, which are cells that send information through the nervous system. “I try to predict how neurons are wired in the brain and understand what kinds of computations those neurons are doing, based on what we know about the behavior of the animal or what we know about the neural response,” she says.

A dragonfly’s reaction time to a maneuvering mosquito, for example, is merely 50 milliseconds. (To compare, a human blink lasts about 300 milliseconds.) Fifty milliseconds is only enough time to cross about three neurons, according to Chance. In other words, to keep up with a dragonfly, an artificial neural network needs to be done processing information after only three steps. But because brains fire many signals at once, each step may involve many calculations running simultaneously.

Missile defense systems today rely on established intercept tech­nologies that are, relatively speaking, computationally intense. Chance is rethinking these strategies via highly efficient dragonflies as a model to see if she can potentially shrink the size, weight, and power (SWaP) needs of onboard computers. Lowering SWaP would enable the development of smaller, lighter, and more maneuverable interceptors and may also reveal new ways to intercept maneuvering targets such as hypersonic weapons, which follow less predictable trajectories than ballistic missiles. Further, it could reveal new ways to home in on a target with less sophisticated sensors than the ones currently used.

Hypersonic weapons pose an existential threat because their blistering speed, altitude, and maneuverability may be capable of defeating most missile defense systems. Russia, China, and the U.S. are all currently working to develop hypersonic weapons or already have this capability, so it will become increasingly important to be able to defend against them.

That said, dragonflies and missiles move at vastly different speeds, so it’s unknown how well this research will ultimately translate to missile defense. But developing a computational model of a dragonfly brain should also provide long-term benefits for machine learning and artificial intelligence (AI). AI – currently used in such disparate areas as the military, self-driving transportation, and prescription drug development – can benefit all kinds of applications that need highly efficient methods for constructing fast solutions to complex problems.

Work is currently ongoing at Sandia to refine Chance’s algorithms to determine where they’re most applicable.

[Figure 1 | SNL scientist Frances Chance is revealing insights into how dragonflies intercept their prey, which might prove useful for missile defense. Photo: Sandia National Laboratories.]